Haemostasis and coagulation

This chapter is relevant to Section Q1(iv) of the 2017 CICM Primary Syllabus, which expects the exam candidates to be able to "describe the process and regulation of haemostasis, coagulation and fibrinolysis." It is an exam topic second only to heparin and antiplatelet agents in its popularity. Past exam questions concerning this area have been detailed, numerous, and difficult. 

• Question 15 from the first paper of 2022
• Question 10 from the first paper of 2019 
• Question 6 from the first paper of 2016 
• Question 20 from the second paper of 2011
• Question 17 from the first paper of 2011
• Question 4(p.2) from the first paper of 2009

In summary:

For the disinterested exam candidate, "Physiological haemostasis" by Simon McRae is an excellent break from Twitter, even more so by virtue of being the ninth chapter of Mechanisms of Vascular Disease (2018) which is, somehow, bafflingly, available in its glorious full form via library.oapen.org. McRae's structured chapter is more than enough to answer the college questions on this topic, but there are numerous others, including Zaidi & Green (2019), Sira & Eyre (2016), Rasche (2001), Palta et al (2014), and probably more.  For the casual reader not preparing for any sort of gruelling barrier assessment, one may also recommend the excellent essay by Spronk et al (2003). Also, overall anything by Maureane Hoffman is excellent, as in general she seems to be the author or co-author of 90% of the papers that were used to put this summary together, as well as the main engine behind major paradigm shifts in the understanding of haemostasis.

Models of haemostasis and anachronisms of terminology

One may come across the term "cell model" when reading about this, often presented alongside the usual "intrinsic/extrinsic pathway" description, which might suggest that there's some kind of controversy raging in the world and haemostasis research, but no - there is not. This term is used to refer to a change in paradigm, which is largely accepted now (over twenty years since it was first proposed). Hoffman & Monroe (2001) challenged the standard model of clot formation (where the factors were regulating the process) by proposing an alternative where the cells were the main regulator and driver. In this cell model, there are several overlapping stages which can all take place simultaneously:

1) initiation, which occurs on a tissue factor bearing cell;

2) amplification, in which platelets and cofactors are activated to set the stage for large scale thrombin generation; and

3) propagation, in which large amounts of thrombin are generated on the platelet surface

This model follows the older "cascade" or "waterfall" model from the 1960s, which focused entirely on the proteolytic enzyme system which progresses as a linear series of reactions (Davie & Ratnoff, 1964). The intrinsic pathway was so called because it was thought not require the participation of substances extrinsic to the blood, whereas the extrinsic pathway required tissue factor. They both converged on a final common pathway which ultimately led to the formation of thrombin and therefore fibrin (as thrombin converts fibrinogen into fibrin). It is now largely agreed that this proteolytic amplification and propagation take place simultaneously with platelet activation, i.e. primary and secondary haemostasis proceeding hand in hand, and one supporting the other. 

However, these "pathways" are still convenient tools for learning about the clotting cascade, and one would be expected to regard them as separate for the purposes of writing an exam answer. In short, to the people who write textbooks, it will be important for you to demonstrate that you've been learning from the textbooks. It appears to be the belief of medical teachers that nobody should be spared the need to learn (perhaps even memorise) the steps on those crazy multicoloured coagulation cascade flowcharts (or monochrome ones; for example, here is one from the 26th edition of "Ganong's Medical Physiology"). Thus, the discussion which follows will attempt to combine the modern way of thinking about haemostasis while still retaining enough of the 1960s terminology to appeal to CICM examiners.

Briefly, about clotting factors

As we will be referring to Roman-numbered factors extensively from here on, it is probably worth spending a moment getting to know them to some limited extent. Their names were decided by a frustrated-sounding International Committee in 1954, which formed and acted in response to the prevailing chaos of the time (at one stage there were fourteen different names in the literature for the same clotting factor). It took these twenty-three scientists four years to come to the conclusion that Factors will be numbered by Roman numerals, in the following order:

• Factor I: fibrinogen
• Factor II: prothrombin
• Factor III: thromboplastin
• Factor IV: calcium (yep, just ionised calcium)
• Factor V: "proaccelerin"
• Factor VII: "proconvertin" 
• Factor VIII: "Antihaemophilic factor"
• Factor IX: Plasma thromboplastin component
• Factor X: Stuart-Prower factor
• Factor XI: Plasma thromboplastin antecendent
• Factor XII: Hageman's factor

The synthesis of Factors II, VII, IX and X requires Vitamin K, which is where warfarin exerts its pressure. As one can see, the half-lives of most of these soluble proteins are relatively short, which means they will soon be depleted if warfarin therapy is commenced. Apart from Factor VIII, the rest of them are synthesised in the liver (and Factor VIII is also mainly synthesised by the liver, but it can also be produced by endothelial cells).

The bar graph below was constructed by a younger, more careless version of the author, and though that guy would not have fabricated data or uncritically accepted untrustworthy resources, he certainly did not record his reference for these values, which means they cannot be verified; and so the graph is included here purely because it looks like it must have taken a long time. 

Of these factors, fibrinogen is present in the blood in the greatest quantities, as it is the final product of the coagulation cascade, and therefore you need a lot of it (whereas in contrast all the other factors play a mainly enzymatic and catalyst role, which means you do not need a lot of them). To coagulate 100 ml of blood you need only about  0.2 mg of Factor VIII, but up to 250 mg of fibrinogen. The normal fibrinogen level is about 2-4g/L, with coagulation becoming affected by levels below 1.0. 

Initiation of haemostasis

Haemostasis is initiated by the exposure of blood to things which blood is never meant to see, which is basically anything that is not normal endothelium. Usually an injury to the vessel wall means that some structural component of the tissues will suddenly protrude rudely into the bloodstream. Though many things can act as the trigger for the process of haemostasis, the most classical is the exposure of tissue factor. The initiation phase of haemostasis is therefore defined as events that take place on or near a tissue factor-expressing surface, such as a fibroblast or vascular smooth muscle cell.

Tissue factor exposure occurs when the endothelium is disrupted

"Tissue factor" is a 45-kDa transmembrane glycoprotein that looks a bit like an immunoglobulin and functions like a high-affinity receptor for Factor VIIa, with which it acts as a co-factor. Structurally, it belongs to the Class II family of cytokine receptors.  It is serenaded beautifully by Grover & Mackman (2018), in case the reader needs a more professional peer-reviewed overview. It was called "tissue factor" mainly because in the early research into coagulation involved starting the process of coagulation by adding some "tissue" to the blood, such as macerated chunks of placenta (Morawitz, 1904). It was only about eighty years later that people were able to purify it. The name persists, even though it is clearly not just in tissues (there's a soluble version), it's not in all tissues (there is less of it in skeletal muscle and joints),  and its role is clearly not just clotting- it appears to be essential to angiogenesis.  When Toomey et al (1996) produced some gene knockout mice who lacked tissue factor expression, all the embryos died around the ninth day of gestation because they failed to develop a normal circulatory system.

Anyway: tissue factor is the main mechanism of initiating coagulation, and it is usually exposed by blood vessel damage because it is expressed at various tissue boundaries, where it forms a sort of "haemostatic envelope" around the circulatory system (Luther et al, 1996). Here's an excellent image from Hoffman, via Mackman (2009), illustrating the expression of tissue factor (stained brown) in the walls of a mouse arteriole:

As you can see, the endothelial layer physically separates this receptor from its circulating ligand. The disruption of the endothelium is therefore the necessary first step to begin the process of haemostasis.

Tissue factor activates the clotting cascade via the "extrinsic pathway"

Say some tissue factor becomes exposed. When tissue factor and Factor VII meet, they quickly bind each other, and activate Factor X, turning it into Factor Xa.  This is all happening in a tiny nanoscale pocket on the surface of the denuded cell that was expressing the tissue factor; Xa can also be rapidly inhibited by "tissue factor pathway inhibitor" (TFPI) or antithrombin III if it moves away from the immediate vicinity. However, if it stays put, it has a chance to combine with enough Factor Va to produce some thrombin. Thrombin is the next major step of the pathway, as it converts fibrinogen into fibrin and activates platelets.

Thrombin activates platelets and creates fibrin

Thrombin is the coagulation pathway enzyme. If one were thrombin, one could flex about the central role one plays in the clotting cascade. Among the many things this serine protease does is cleaving fibrinopeptides A and B from the α and β chains of the fibrinogen molecule, making it spontaneously polymerise into insoluble fibrin polymer strands. That, reader, is one of the victory conditions of the clotting cascade. Check and mate.

However, fibrin alone is not enough to create a clot with good integrity. For this reason, thrombin also activates platelets. Not a lot is required: Monroe et al (1996) were able to demonstrate platelet activation with a concentration of just 0.1nM. Platelet activation occurs when thrombin binds to a protease-activated receptor (PAR) on the platelet surface, which are G-protein coupled receptors (Hung et al, 1992). This step is a big deal. To paraphrase from Hoffman (2001), the rest of the action moves from the tissue factor bearing surface, and onto the surface of the platelet. 

Other things also activate platelets

Because it would not be logical to put everything into one basket, the process of primary haemostasis can be initiated by all sorts of platelet-activating mechanisms, which also need to be mentioned at this stage. In fact it would have been good to mention them at an even earlier stage, as all of them are available at the same time as the nude tissue factor starts to do its thing with Factor VII. However, the need to present this information in some sort of logical sequence has resulted in a separation of otherwise concurrent events. So: numerous other things can cause platelet activation, and in case the reader needs any more detail than what is available in the crude diagram below, they are invited to review Saboor et al (2013), whose coverage is exhaustive.

Stalker et al (2016) expands on the intracellular pathways via which these receptors tend to produce their subsequent effects. In short, they are all G-protein coupled receptors, except for the receptor which binds collagen (it is linked to tyrosine kinases). That is probably not important. The most essential part of this to remember is that all of these simultaneously activate platelets.  

Platelet adhesion, activation and aggregation

So: at the same time as tissue factor is producing some early thrombin deposits on the denuded injured surface of the defect in the blood vessel, the platelets are adhering, activating and aggregating. What exactly is meant by these terms?

• Platelet adhesion is mainly mediated by extracellular matrix proteins exposed by the removal of vascular endothelium. The most important of these is probably collagen, as the collagen receptor (α2β1 integrin, a different receptor to the one which causes platelet activation) is apparently the most numerous surface receptor on platelets  (Ruggeri et al, 2007, give a figure of 80,000 copies per platelet, which is apparently a lot). Exposed surfaces denuded of endothelium are going to be bristling with collagen, and platelets bind to them. Collagen is not the only protein that does this: other thrombogenic proteins include fibronectin, laminin, fibulin, vitronectin, Von Willebrand Factor (which in turn binds to collagen anyway) and probably many others. The interaction is so straightforward that it probably did not require a diagram, but here we are:
  
• Platelet aggregation is mediated by fibrinogen and fibrin via the GPIIb/IIIa receptor. This receptor is an integrin that can bind to both soluble fibrinogen, fibrin and to Von Willebrand's Factor. For the pragmatic intensivist, this step is worth knowing about because it is the target of GPIIb/IIIa antagonists such as abciximab and tirofiban. 
  
• Platelet activation is a metabolic process that is triggered by the aforementioned receptors binding their ligands (PAF, ADP, Von Willebrand Factor, collagen, thromboxane A2, etc). All of these G-protein coupled receptors ultimately do the same molecular thing, which ultimately leads to the increase in intracellular cAMP in the platelet, and the increased availability of intracellular platelet calcium. There are three main results from this activation:
  • a change in the shape of the platelet,
  • the release of stored mediators from granules, and most importantly
  • phosphatidylserine exposure on the platelet surface, which allows more clotting factors to bind there

Released mediators unleashed by activated platelets include more clotting factors (Fvvactor V), substances which activate more platelets (eg. ADP and thromboxane A2) as well as locally acting mediators which influence the regional circulation (eg. serotonin). Local vascular smooth muscle constriction results. Chen and Tsai reported on this in detail in 1948, though earlier researchers had noted how there was a tendency for the distal segment of an injured artery to become pulseless. Though the photography techniques of the 1940s was probably too poor to capture the rabbit's ear arteriole in truly glorious detail, the investigators used a camera lucida to trace the outline of the vessel, and observed that it began to constrict immediately and was completely obliterated within 30 seconds of the puncture. The contracted vessel began to relax 8-15 min. later, but did not regain its former size until 10-20 min afterwards (if the clot was washed off), or up to an hour if the clot remained in situ. 

The shape change which platelets undergo is remarkable. It goes from boring biscuit to shoggoth.  Nothing would describe it better than this image from Aslan (2017):

This "sequence" of actually simultaneous events triggered by injury (tissue factor medicated activation of the "extrinsic" pathway of the clotting cascade, generation of thrombin, early formation of some fibrin and the activation of platelets with local mediator release) seems to be the defining features of "primary haemostasis". The term itself does not have a clear definition, and even authors who use it in their title often do not make an effort to clarify exactly what they mean by it (looking at you, Berndt et al, 2014). From perusing a selection of both professional and apocryphal materials, the word "platelet plug" appears to be used a lot, suggesting that platelet aggregation is the dominant feature. "Secondary haemostasis", in contrast, tends to be used in reference to events involving the intrinsic and common pathway, as well as the contraction of the maturing clot.

Secondary haemostasis,  intrinsic and common pathways

The term "secondary haemostasis"s corresponds to the amplification and propagation phase of the cell model. Again, the term does not seem to have a fixed definition, but in the literature it seems to be associated with descriptions of the intrinsic and common pathway, as well as their disorders (for example, haemophilia is thought to be a defect of secondary haemostasis). A representative description from McRae:

"Secondary haemostasis describes the process whereby exposure of tissue factor to the bloodstream leads to a series of enzymatic reactions that result in a sufficient burst of thrombin production to convert soluble fibrinogen into a stable network."

Amplification and the intrinsic pathway

The last stages of the primary haemostasis process, where the platelets are activated and covered in negatively charged phosphatidylserine, sets the scene for the part of the haemostatic process where large amounts of thrombin can be generated in order to produce an even larger amount of fibrin. To produce thrombin on an industrial scale, a lot of Factor V and Factor VIII will be required. Fortunately, Factor V comes from platelet granules, and Factor VIII comes with Von Willebrand Factor, because they circulate as a complex. Thrombin cleaves Factor VIII off VWF and converts both FVIII and FV into their active forms, which bind to the platelet surface.  This is the necessary step to set up the conditions for a "thrombin burst", as  FVIIIa and FVa can then activate more FX via the intrinsic pathway, and the FXa can activate more thrombin in a self-amplifying positive feedback loop.

Propagation and the common pathway

The intrinsic pathway, as it is usually depicted, starts with the activation of prothrombin (Factor II) into thrombin (Factor IIa). As has already been abundantly mentioned, this whole thing is happening simultaneously, and there is enough thrombin around to do this. Thrombin activates Factor XI, which in turn activates IX, and IXa acts as a cofactor with FVIIIa to produce more FXa, and therefore more thrombin via FXa and FVa.  

Now, from this positive feedback loop, one might expect a vast amount of thrombin to be generated, and this is indeed what happens. The result of this thrombin excess is a large-scale industrial conversion of soluble fibrinogen into polymerised crosslinked fibrin, which increases the stability of the clot. This last phase, where Xa and Va produce thrombin which then generates fibrin, is generally referred to as the common pathway, as both the intrinsic and extrinsic pathways converge here.